A Measurement Software for Professional Training in Early Detection of Melanoma

Comments and Suggestions for Authors

Table 1 is badly formatted, what makes it hard to read (consider transposing it, and/or changing font size to smaller).

Thanks for your comments. We have appreciated, table has been formatted again

Pay more attention to uniform formula notation within the text (inline) and formulas, e.g. you are missing italics and lower indexes in the inline formulas and symbols (e.g. line 131 - Fi(x), while in (1) it is F_i(x), and in Fj(x), j is not an index, even in formula 1.

Thanks for your comments. We have appreciated, formulas have been edited

Don't understand the notation in formula 4 - what is the argument, and what is the equation, that is being maximized?

Thanks for your comments. Formula has been edited

Table 5 - don't define measures within the table column. I think it would be better, to show data characteristics (class count) in one table, and the classification outcome in second table (only in percentage). Does it have any point to show diagnostic and clinical accuracy in tables, when they do not differ?

Thanks for your comments. We have appreciated, table has been formatted again, results has been checked, clinical accuracy is similar to diagnostic accuracy and has been reported in order to show the different performance of two groups of dermatologists when enfaced with the problem of the lesion excision (that is the matter for the patient, since the melanoma diagnosis is formulated and communicated only after biopsy analysis)  

This is definitely not the only one automated medical decision support system for image based diagnosing of melanoma, so I'm missing a reference/comparison to the efficacy of concurrent systems.

A comparison of the software system to the efficacy on concurrent systems has not been reported because the latter group typically includes proprietary systems. When sufficient details about implementation are provided, a common database is missing to check them. On the other hand, the goal of the paper is not to show the developed software outperforms the digital tools proposed in literature (it exhibits similar sensitivity and specificity as discussed in the final section where some references to systematic reviews are reported). Instead, the goal is to show the importance to develop automatic systems according to well-known diagnostic methods so that the final users (dermatologists) are able to understand the software results and consequently improve the own clinical decision.

Comments and Suggestions for Authors

Cacciapuoti et al describe a potential developed measurement software to detect melanoma. This is an interesting and important research article since the importance of early detection of the disease is highly needed due to the high mortality. Overall, the manuscript is well written and easy to read. The authors suggested crucial 7 point check list for the detection and comprehensive algorithm as a part of the software development. Therefore, I would like to recommend this article to be published in Applied Sciences journal. There are several points that could be addressed by the authors to improve the manuscript:

Thanks for your comments. We have appreciated.

1.Should include melanoma in the title since this is the major aspect of the article.

Thanks for your suggestion. Title has been modified in “A measurement software for the professional training about early detection of melanoma”

2.How does the software could differentiate the detection of Non-melanoma skin cancers such as SCC and BCC?

The developed software system implements the 7-Point Check List because is intended to support the early diagnosis of melanoma that often occurs correspondingly to pigmented melanocytic lesions.

Indeed, the Step 1 of dermoscopic analysis is intended to differentiate the melanocytic lesions from other (mainly benign) types of skin cancers such as Squamous Carcinoma Cell (SCC), Basal Carcinoma Cell (BCC), Angioma, Seborrheic Keratoses (SK).

At Step 2 (limited to the analysis of melanocytic lesions) dermatologists typically adopts a diagnostic method such as the 7-Point Check List to classify the pigmented lesion as melanoma (that is the most dangerous skin cancer).

Thus, the software could be adopted for determining such dermoscopic features also present in not melanocytic lesions such as SCC and BCC (typically Arborizing Vessels similar to AVP or  Blue-gray Ovoid Nets similar to Blue Veil), but only after developing a suitable classification layer for Step 1, that is beyond the actual scope of the present automatic system.

3.Could the expert group include the real physicians instead of senior researchers just to be consistent with the non-expert group which also consists of 3 physicians with less experience?

Thanks for your suggestion. In truth, the expert group already includes 3 real physicians. They have been indicated as researchers just because they serve at Public Hospitals that are also Academic Departments (including University of Federico II of Naples and University of Salerno). Thus, the text (line 249) has been revised.

4.In the outlook, could this software be developed to be a user-friendly for common people?

The software has been developed as a digital tool supporting the training of physicians during the post-graduated courses in dermatology and corresponding clinical activity in order to increase the diagnostic performance (in terms of sensitivity) about the early detection of cutaneous melanoma when suspicious pigmented lesions are to be examined. To pursue this goal the software system takes into account dermoscopic images and is mainly based on 7-Point Check List. As outlook, in order to boost massive screening campaign, the results of the automatic image processing could be adopted as dermoscopic triage carried out by general practitioners and/or pharmacists, but it should be never used by common people for self-analysis of cutaneous lesions.

Thanks for your suggestion, it has lead us to include the above discussion in the section Featured Application that has been added to the new version of the paper